Satellite real-time guidance task planning method and system based on deep reinforcement learning

Abstract

The invention provides a satellite real-time guidance task planning method and system based on deep reinforcement learning. The method comprises a scene establishment step: establishing a'time attitude 'two-dimensional satellite imaging task planning training scene; an extraction and conversion step: extracting and converting agent training element states, actions, costs, rewards and ending conditions of the satellite imaging task planning training scene; a docking step: integrating with the satellite imaging task planning training scene by using a deep learning algorithm; and a reinforcementlearning step: learning the satellite imaging task planning process by using deep reinforcement learning. The satellite orbit and the longitude and latitude of the target are subjected to'time side-sway 'two-dimensional mapping, the dimensionality of a reinforcement learning environment state space is reduced on the premise of not sacrificing data precision, and the convergence speed of intelligent agent training can be increased on the premise of reserving all effective information.

Description

technical field [0001] The present invention relates to the field of satellite mission planning, in particular to a method and system for satellite real-time guidance mission planning based on deep reinforcement learning, especially a method for implementing deep reinforcement learning training for remote sensing satellite real-time guidance mission planning. Background technique [0002] In the current remote sensing satellite field, there is a contradiction between limited observation resources and increasingly complex and real-time imaging requirements. Improving the scheduling level of observation resources can make limited observation resources better adapt to complex and time-sensitive task requirements. This makes satellite mission planning a research hotspot. [0003] In recent years, the field of imaging satellites has begun to use a real-time guided imaging system that uses discovery payloads and confirmation payloads to cooperate with target discovery and imaging....

AI Technical Summary

Problems solved by technology

Compared with the published methods: Wang Chong, Research on Agent-based Distributed Cooperative Mission Planning for Earth Observation Satellites (Ph. There is a longer learning period

Images